use two numbers to describe a printer抯 resolution, one for the vertical scanning and the other for horizontal. You can use both of these numbers to get a range of the dpi which can be used for high definition printing. For example, a printer with a resolution of 1200 x 600 will print a high definition image at 400 to 200 dpi. At 400 dpi, the image will be printed smaller and sharper. For many modern color printers, experts recommend 300 dpi as being a reasonable number (Milburn, 2000, p. 31).

The resolution of a file can be changed up or down by resizing the image. However, if the size of an image is increased or scaled up, the space between the original pixels must be filled in. The software used to make the file larger will fill in the space between the original pixels, however the result is often very fuzzy looking. Newer imaging software uses the process of interpolation which is a more intelligent way of choosing the pixels which are used as filler, but in general images which are scaled up more than ten or twenty percent lose much of their sharpness. Higher end imaging software programs use more sophisticated interpolation algorithms and tend to produce better results. Reducing the resolution of an image can generally be done to a much greater extent with no noticeable loss of quality. In order to reduce the resolution of an image, the imaging software must throw away pixels. Images can often be scaled down by more than fifty percent and still be of acceptable quality.

Resolution describes the amount of detail in an image, but the tonal quality of each pixel must also be considered. The pixel depth, also known as bit depth of an image, controls its tonal qualities. Milburn (2000, p.21) describes Apixel depth@ as bits per pixel because the amount of color information which can be assigned to each pixel depends on the amount of data that can be assigned to that pixel. A 1-bit image can only display black and white. To calculate the number of colors which can be displayed in an image with a particular bit depth, Murray & VanRyper (1994, p.861) provide a formula B The number 2 raised to the power bit depth specifies the maximum number of values the pixel can assume.

For example, an 8-bit image can display 256 colors because 2 to the 8th power is 256.

One of the basic facts of digital imaging, with many practical applications, is that the resolution and bit depth of an image affect the physical dimensions, file size, and tonal quality. In a simple example of how this information can be used, let us consider pixel depth and its effect on tonal quality: The human eye can only recognize 256 shades of gray, so an 8-bit image is adequate to capture black and white photography.


Many users who are not involved in image-processing will also find 8-bit images acceptable for viewing photographic images in tasks involving a web browser. Many IBM compatible computers in the early 1990s used primarily for office applications such as word processors and spreadsheets, which generally contained no photographic images, were originally set-up to display only 4-bit images, a maximum of 16 colors, even though the hardware and software were capable of displaying 8-bit images or better. Users of typical office applications rarely complained about this; however when the web became common in the workplace, these computers required an adjustment of the display software so that they could display at least 256 colors (8-bit) and output photographic images of acceptable quality.

Another important aspect of a digital image is its file type. A variety of formats have been devised for image data because each type meets different types of images and different end uses. A primary consideration is which file type can produce the smallest file size in bytes. This is important because smaller files are cheaper to store and can be transmitted more quickly across a network. Three of the most common image file types are GIF, JPEG, and TIFF. GIF and JPEG files have built in compression schemes that control the size of the files. TIFF files can be either compressed or uncompressed.

The JPEG file type is designed for natural, real?world scenes such as those found in photographs, realistic artwork, and similar material. JPEG, which can contain up to 24 bit color (16 million colors) uses a Alossy@ compression scheme which throws away some of the data captured in the original subject. This compression scheme is based on the knowledge that Athe human eye is limited in the number of colors which it can distinguish simultaneously particularly if those colors are not immediately adjacent in the image or are sharply contrasting.@ (Murray & VanRyper, 1994). The JPEG format can produce images which are drastically reduced in file size, yet the removal of color information may not be easily noticeable.

The GIF file type can only contain 8 bits of data (256 colors) and is a better choice for lettering, simple cartoons, line drawings or subjects that contains large expanses of single colors. Many well-designed web sites take advantage of these formats by using GIFs for buttons and other simple graphics, and JPEGs for photographic images.

The TIFF file type is often used by professional photographers, graphic artists, and others because of its numerous features including the ability to store files of any pixel depth that are compressed or uncompressed. TIFFs can contain textual data (metadata) about the image that is not visible in a pictorial display of the image, but can be viewed with appropriate software. TIFF is widely used by museums and other institutions as their primary archival format. A newer format, PNG, like TIFF, is very full featured, and is beginning to be used both as a format for the web and archiving. Most popular image editing programs and recent web browsers, including the 4.0 versions or later versions of both Internet Explorer and Netscape Navigator, can handle PNG files.

Image-processing Software


There are many reasons for editing a digital image. One of the most common problems with a digital image is that its resolution is much too high for use on the World Wide Web. The low resolution of computer monitors necessitates the resizing of high-resolution images created by digital cameras and scanners. These images are well suited to printing at 200-300 dpi in order to produce a high If one tries to use a web browser to view an 2000 pixel wide by 1500 pixel high image produced by a 3 megapixel camera on a typical computer monitor, usually set to 800 x 600 or 1024 x 768, only a small portion of the image will be visible. These high-resolution images originally intended to be used in print publications or for archival purposes need to be scaled down for use on the Web. Images intended for print or archival storage in TIFF format or in some other file type that cannot be viewed on a web browser need to be converted to JPEG or GIF format. Large image files being prepared for the web need to be compressed, and perhaps have their pixel depth reduced in a process that is referred to as Aweb optimization. @ Also, common graphics such as icons which are used repeatedly on a web site should use the Aweb safe palette@ in order to have their site look consistent across different computer platforms. This palette refers to the 216 colors that are shared by both Macintosh and Windows computers when using a web browser.

There are some common editing processes that are done to images of all kinds regardless of whether they are intended for the web or for print publication. The color, contrast, and brightness of an image may need to be adjusted. Photographic images of people and animals often require Ared eye reduction. @ When only a small portion of an image is needed, the image needs to be Acropped@. This involves selecting and saving a rectangular portion of the image. Images scanned from books or journals often require that the margins be cropped since they usually contains no useful information. Large collections that are being processed for use on the web may require the same repetitive tasks to be done to thousands of files in what is known as batch processing.

Because of enormous competition in the imaging software industry, publishers are constantly adding new features to their programs, so in a very short period of time programs can change dramatically. However, it is worthwhile to look at some categories that exist at the end of the year 2000. Many criteria mentioned can also be used to review other image-processing software.

Professional image-processing programs that have a wide range of features are commonly used in office settings, and as their prices drop, are becoming more popular. These professional image-processing programs can easily perform all of the basic processes such as cropping, resizing, and color correction, as well as more advanced processes such as the use of image layers, which go beyond the scope of this paper. Because of intense competition among software publishers in recent years, many of the less expensive products (under 50 dollars) have added features that were once only available on high-end packages.


For the purpose of this paper, some rough separators between the software intended for professionals and those intended for home use is that the professional software tends to have many more image-processing tools, support more file types, and provide high end features such as layering and digital water-marking. Another feature that is available in most of the professional programs is the ability to use pressure sensitive Wacom tablets. Also the icons that represent the tools provided in these professional packages are often arranged in densely packed toolbars. This provides more speed and power, but is a harder to learn than imaging software for home use, which offer a less dense layout without so many icons to memorize. Also, imaging packages for home use often have Awizards, @ which help a user to accomplish a particular task in a simple (slow) step-by-step fashion. Low-end image-processing packages intended for home use will not be covered in this paper.

Adobe Photoshop is the most well known of the professional image-processing programs, and many of its competitors are trying to add all of the features of Photoshop to their products. Photoshop grew out of the (paper) publishing industry, and is widely used today in the preparation of images for print publications and for the Web. Photoshop=s competitors try to match it=s features and to be compatible with Photoshop Plug-ins. These Plug-ins are small programs, which are added to Photoshop to increase its functionality. An important new feature of Photoshop 6 and of some of its competitors is the ability to create and edit vector graphics that was, until recently, only the domain of drawing programs such as CorelDraw.

Some of Photoshop抯= better known competitors include Corel PhotoPaint, PaintShop Pro, and Live Picture. These programs are all available at a lower cost than Photoshop, and PhotoPaint comes bundled with the well-know vector graphics program, CorelDraw. Some of them such as PaintShop Pro seem to emulate the interface of Photoshop. Another notable features of PaintShop Pro is its ability to open very rapidly. When only a single image needs to be quickly cropped or resized, this ability to open the program quickly, complete your work, and be done, is appreciated. PaintShop Pro has image-management features that will be discussed in more detail later.

Macintosh and Windows computers can both run Photoshop, but Unix/Linux computers cannot. A popular image-processing program in the Unix/Linux world for several years, and recently available for Windows, is the (free) open-source program, AThe Gimp, @ often referred to as a Photoshop clone. Besides its Photoshop-like interface,and numerous features, the Gimp has a powerful scripting language that allows very sophisticated batch processing. Other well-known image-processing tools in the UNIX world are xv, and Image Magick, but these lack the power and refinement of the Gimp. The CorelDraw Suite, which includes CorelDraw and PhotoPaint, is now available for Linux.

Imaging software intended almost exclusively for web development has become popular in recent years. The most well known of these programs include Macromedia Fireworks, Adobe ImageReady, Ulead PhotoImpact, and Ulead Web Razor. Any of the professional all purpose packages, such as Photoshop, can be used to do tasks such as optimizing files for the web by reducing their pixel depth; however many users have found these programs dedicated to preparing files for the web to be particularly effective at these tasks.


Batch processing, the ability to perform repetitive tasks on a group of files, is a very important function of imaging programs. Many programs have this feature, including Adobe Photoshop which has the ability to record processing steps which are performed on an image using a tool called AActions@. These steps are saved and can be performed on a single image or a batch. Corel PhotoPaint and several other professional programs have this feature. Equilibrium software=s DeBabelizer Pro is a well-known program for Mac and Windows, specializing in batch processing. DeBabelizer has a graphical interface, but also uses scripting.

In the Unix/Linux environment, the Gimp provides batch processing ability through the use of AScript-Fu, @ a full-featured scripting language geared toward programmers. An older tool that works well with only a command line interface, NetPBM, is an open source toolkit, consisting of 180 separate small programs. Used primarily for the conversion of images between a huge variety of different formats, NetPBM also allows some basic image operations such as resizing, and changes in pixel depth. This toolkit allows users with only a very basic knowledge of Unix and little or no programming experience to perform powerful batch processing from a command line.

Specialized 搒titching” programs are used to automatically stitch together smaller images that have some overlap in order to create a large image. These programs can stitch together images horizontally to create panoramas, and some can stitch together images arranged vertically and horizontally. They include QuickStitch, Photovista, and Visual Stitcher. Canon PhotoStitch can stitch together images to create a 360-degree panorama that is rotated with a java applet. Some all-purpose image-processing programs, including MGI PhotoSuite, also contain stitching features.

Another specialized program, PhotoGenetics is a unique Aimage adjustment@ program which has become popular among professional and amateur digital photographers.

ATo improve an image using PhotoGenetics, a person rates the original against variations produced by the software. Simply by clicking on a scaled bar, the preferred enhancements transfer to the original image. When the user is satisfied with the results, the new image is saved. The set of enhancements may also be saved separately and applied quickly to similar images in the future.” (QBeo, Inc., 2000)

A final specialized type of image-editing program is one that can edit the EXIF data, textual information contained within a JPEG image. Most digital cameras automatically record the shutter speed, aperture and other data in the EXIF field. EXIFREAD, and Thumber are two programs that allow this data to be extracted and edited.

Image Management Software

Image management software ranges from simple freeware tools which allow you to browse thumbnail image tools to huge client-server based AMAM@ (Media Asset Management) systems which integrate with Oracle or MS SQL server. MAM systems from Quark, Canto and Extensis are widely used by corporations, and large government institutions. These systems are beyond the scope of the paper.

Perhaps the most basic, and also the most useful image management feature is the ability to rapidly browse through a directory of thumbnails, small images which are miniature versions of larger images. These thumbnails appear as a grid on your screen and are linked to the larger versions of the images. This allows the original images to quickly be located, viewed, edited, or deleted, even if no special database system with textual data was used to store them, but it does help if the directory that holds the images has a descriptive name, such as ANew York City - Times Square January 2000. @ Photoshop 6.0 does not have the ability to allow the user to browse a screen of thumbnails linked to the original images, but Paintshop Pro, Corel PhotoPaint, Ulead Photoexplorer, and inexpensive shareware programs such as Thumbs Plus all provide this extremely useful feature.

The next step up in image-processing programs are those that incorporate some sort of searchable database system which allows the user to attach text to individual images and directories. This text can then be searched. Extensis Portfolio, single user version, is an inexpensive program which contains this feature. Individual 揷atalogs,@ groups of images, can be posted on the web along with the free catalog viewer which allows users to search the text of the catalog. The catalogs can be read by both Mac and Windows computers (Extensis Products Group, Inc., 2000). Canto Cumulus is a bit more expensive, but has similar capabilities.

Another feature of some image management programs is integration with the web. At the lowest level, this can include the automated creation of HTML from a page of thumbnail images. Thumber, Thumb Plus and others have this capability. (Curtin, 2000) The resultant HTML page consists of a grid of thumbnail images with links to larger images. Some programs such as Thumber, have t

he option of outputting EXIF data such as the shutter speed and aperture under each image. Many of the higher end software producers, including Canto, Extensis and Quark others also sell special web server software that allows users to search textual data about the images. They also allow developers to upload images. The simplest of these image database/web servers are only intended for small workgroups, but the higher end packages are integrated with client-server databases such as Oracle and MS SQL Server. These high end packages can often be run on either Unix or Windows servers. Filemaker Pro, the inexpensive database package, has a built-in web server that can display thumbnails that are hyperlinked to the larger versions. Filemaker Pro is well suited to small workgroups sharing images. Filemaker also has a higher end web database server version. Filemaker is available for Windows and Mac.

Conclusion


A wide range of options exist for image editing and image management. Even professional level software is reasonably priced, especially for those who are receiving the standard educational discount provided by many major software publishers. Because of this low cost it is often wise to try a range of programs to find the one which best meets your needs or the needs of a particular project. Any of the packages mentioned in this paper can easily be located by doing a keyword search on one of the Web search engines such as www.google.com or www.altavista.com. More detailed explanations of many of the concepts can also be found in this way. The specifics of the software packages mentioned will change rapidly as the industry evolves, but it is hoped that the general concepts about digital imaging which were covered and the overview of some of the packages can be helpful, both now and in the future.


References

Curtin, D. (2000). The digital darkroom: Image management software. Boston, MA.: ShortCourses Publishing Company. http://www.shortcourses.com/book03/07.htm

Extensis Products Group, Inc. (2000). Entensis portfolio 5. Portland, OR: the company. http://www.extensis.com/portfolio/

Milburn, K. (2000). Digital photography bible. Foster City, CA: IDG Books Worldwide.

Murray, J.D., & VanRyper, W. (1994). Encyclopedia of graphics file formats. Sebastopol, CA: O=Reilly & Associates.

QBeo, Inc. (2000). PhotoGenetics at a glance. North Bend, Washington. http://www.qbeo.com/en/photo_gen_facts.html